Four color continuous ink jet printers produce a continuous stream of ink droplets which are selectively placed onto a printing medium, such as a piece of paper, or deflected into a waste system. It will be appreciated that the term "droplet" refers to a unit of colorant.
When one or more droplets of one color are placed on a piece of paper in an addressable location, known as a "pixel" or picture element, the sum of these droplets are referred to as a "dot". A pixel location can have up to four dots, which include the three subtractive primaries (eg. cyan, magenta and yellow) and black. The amount of colorant at a pixel location can be varied by altering the number of droplets in a dot and the number of dots in a pixel.
In addition to varying the number of droplets in a dot, the positioning of dots on a printing medium can also increase the range of colors printed. This can be achieved by placing dots in a matrix of pixels referred to as a "macropixel." The macropixel is typically a square multiplicity of pixels (e.g. 2.times.2, 4.times.4) or alternatively, it can be a rectangular multiplicity of pixels (e.g. 2.times.4).
Using dots with different amounts of droplets and a macropixel matrix, a macropixel filling method can be designed which determines the dot size and the dot positioning required to produce various shades (i.e. grey levels) of a given color.
U.S. Pat. No. 4,367,482 to Heinzl describes a method and apparatus for representing polychromatic half-tone images employing the formation of image spots (i.e. dots) of equal size.
The U.S. Pat. No. 3,977,007 to Berry et al. describes a monochrome method for placing the dots on the paper utilizing a 4.times.4 matrix (i.e. 16 pixel locations) for each macropixel. Sixteen incremental macropixel fill patterns are defined, defining 16 grey levels. Darker colors are produced by increasing the number of droplets per dot at each pixel location.
U.S. Pat. Nos. 3,977,007 and 4,065,773 to Berry both describe a method for generating multiple grey levels using a combination of dot size variations (i.e. different number of droplets per dot) and dot placement matrices (i.e. different macropixels).
The locations in a 4.times.4 macropixel are typically numbered as follows:
______________________________________ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ______________________________________
where the raster line or printing line moves horizontally across the macropixel matrix. Thus, the dots 1, 2, 3 and 4 are printed first and the other rows follow.
A conventional macropixel filling method for a 4.times.4 macropixel follows the form of a Bayer, or dispersed dot, pattern which is typically used for output devices which print discrete dot sizes. An example pattern is:
______________________________________ 1 9 3 11 13 5 15 7 4 12 2 10 16 8 14 6 ______________________________________
where the macropixel is filled in the order indicated. In other words, a 4.times.4 macropixel having a grey level of 3 would have the pixel locations marked 1-3 filled with a dot and a macropixel with a grey level of 4 would have the pixel locations marked 1-4 filled with a dot.
In direct digital color printing, such as dye sublimation printing and ink jet printing, the conventional technology exemplified by the prior art described above is unable to provide consistently uniform prints without visible patterning over all input density ranges. A typical image which displays an image defect known as rainbowing, is illustrated in FIG. 1A. A simplified enlarged illustration of parts of the image of FIG. 1A appears in FIG. 1B.
The methods of Berry and the pattern of Bayer are not intended for use in color printing. These methods and the pattern would be suitable for use in printing more than one color, were the dots of the different colors always precisely placed `dot-on dot`, as shown at reference numeral 9 in FIG. 1B, for a plurality of black dots 10 and a plurality of other dots 12 of either cyan, magenta or yellow. One black dot 10 and one other dot 12 together form a pixel 14.
In the dots illustrated at reference numeral 9 in FIG. 1B, the dot size is such that there exists some white space 16 surrounding the dots 10 and 12. In this situation, the density of the black dot 10 partially or completely hides the color of the other dot 12 and the amount of white space between each pixel 14.
Unfortunately, as is known in the art, the accuracy of ink jet printer color registration is such that the dots of the two or more colors are not placed consistently so as to perfectly overlap one another. This is shown in FIG. 1B at reference numeral 11, and results in a color which is not the desired color, but rather, one close to it. In this situation, the color to color registration is not dot-on-dot, the density of the black dot 10 does not partially or completely hide the color of the other dot 12 and the amount of white space 16 between the pixels 14 is reduced.
The color defects produced by the misregistration shown in FIG. 1B may be classified as follows:
a) Color variations across the color print which occur at frequencies of 1 cycle/inch or higher (Type I defect or "striations"); PA1 b) Color variations across the color print which occur at frequencies lower than 1 cycle/inch (Type II defect or "rainbowing"); PA1 c) Color variations between two or more color prints of the same image on a single printer (Type III defect); and PA1 d) Color variations between color prints of the same image on two or more printers (Type IV defect).
It will be appreciated that types I and II produce one-dimensional moire patterns which are noticeable in the color print, as seen in FIG. 1A. Generally, the defects are only produced with colors ranging from those which have white highlights (i.e. white spaces 16) to midtones. The defects are not noticeable when the printing dot color is dark enough that the entire printing dot space is covered with ink.
Color to color misregistration is typically caused by inaccuracies in the electro-mechanical design and/or manufacture of the printer. One particular cause is the inaccuracy of the gearbox and drive mechanism, including pulleys, belts, worm gears and lead screws. The variation in the gear operation will produce different frequencies of rainbowing.
While the misregistration can be improved by improving the physical aspects of the printer, this is typically costly, both in the design and the manufacturing processes.